A Comparison of the Fixed and Floating Building Block Representation in the Genetic Algorithm

Abstract

This article compares the traditional, fixed problem representation style of a genetic algorithm (GA) with a new floating representation in which the building blocks of a problem are not fixed at specific locations on the individuals of the population. In addition, the effects of noncoding segments on both of these representations is studied. Noncoding segments are a computational model of noncoding deoxyribonucleic acid, and floating building blocks mimic the location independence of genes. The fact that these structures are prevalent in natural genetic systems suggests that they may provide some advantages to the evolutionary process. Our results show that there is a significant difference in how GAs solve a problem in the fixed and floating representations. Genetic algorithms are able to maintain a more diverse population with the floating representation. The combination of noncoding segments and floating building blocks appears to encourage a GA to take advantage of its parallel search and recombination abilities.

An incremental genetic algorithm approach to multiprocessor scheduling

“…Location independent problem representations, where the information content is not fixed at specific locations on a GA individual, have been proposed in a number of studies as a way to help a GA identify and maintain tightly linked building blocks. Such representations allow for rearrangement of encoded information [7], [5], [16], [19], [23], [35], [40], [46], [48], overlapping encodings which can be more space efficient [7], [42], [45], and the appearance of noncoding regions which affects crossover probability [7], [32], [15], [33], [38], [43], [46], [47]. In some location independent representations, the arrangement of encoded information will determine what is expressed [19], [23], [35] even though the actual encoded content is not determined by its location.…”

Emergence of genomic self-similarity in location independent representations

“…Numerous studies have focused on representations that allow a GA to dynamically evolve the arrangement of information on an individual including the arrangement of individual characters [1,9,13,14,16,21] and the arrangement of groups of characters [5,29,36,42]. While these approaches allow for dynamic evolution of the organization of the encoded information, the mapping from a GA individual to a problem solution still retains some aspects of order: (1) the order in which characters appear in an individual may affect their expression, and (2) some sort of reordering is typically required to decode an individual into a solution.…”

Winnowing Wheat from Chaff: The Chunking GA

“…Memory is possible because non-expressed information is not subject to GA selection pressure making it possible to retain information that may not be currently optimal. Examples of redundant representations include diploidy and dominance mapping [14][15] [16] [17], the messy GA [9], the structured GA [18], and the floating representation [19]. Additional studies on the dynamics of redundant representations include [20][21] [22] [23].…”